Procedural generation of angry birds fun levels using pattern-struct and preset-model
Yuxuan Jiang, Tomohiro Harada, Ruck Thawonmas · 2017
In this paper, we describe our design of a generator for automatically generating fun levels in the famous game Angry Birds and its clone games. In particular, we aim at generation of character-sequence levels. In order to generate fun levels, we propose a pattern-struct approach and a preset- model approach. The pattern-struct approach generates each time different appearance for one of the alphabetical letters, numeral numbers, or some symbols. The preset-model approach is used for increasing the legibility and diversity, where models are built in advance for some capital letters and numbers. Our level generator, called Funny Quotes, was the winner of the Fun Track of the CIG 2016 Level Generation Competition held in 2016 IEEE Conference on Computational Intelligence and Games. After raising issues residing in Funny Quotes through user-evaluation results in terms of legibility of characters in generated levels and of playability of those levels, we propose and evaluate new mechanisms for ensuring the disparity in difficulty among easy, normal, and hard levels for the 2017 competition.